The overall goal of this work is to understand how the motor cortex constructs movement plans and implements them as actions. Previous work in this project demonstrated that neural population activity in the macaque monkey primary motor cortex (MI) cortex predicts hand trajectory during reaching movements. For MI to direct reaching towards a goal, spatial information about the body and the world must be transformed into a motcr framework. Parietal cortex is likely to provide this spatial framework, but little is known about front oparietal interactions. Parietal area 5d provides direct input to MI and is one likely source of arm motion, gaze, and visual target information. The proposed studies will use 100 electrode recording arrays chronically implanted into 5d and MI to study of neural encoding and interactions during continuous tracking of a randomly moving visual target (CRT task). This novel recording method allows the direct and simultaneous examination of neurons in two interconnected areas under identical behavioral conditions. The new CRT task provides a statistical basis for spike and population analysis of these neurons. AIM 1 experiments will compare 5d and MI neurons during CRT. Motor, gaze and visual signal spatiotemporal receptive fields will be characterized and neural information content will be evaluated by reconstructing hand motion, eye position and visual signal from population activity. Encoding in each area will be compared to determine the form and time course of information exchanged. AIM 2 experiments will determine whether 5d- MI networks encode movement in an abstract spatial or body-part based framework. These experiments will further test whether information is in hand, joint or muscle coordinates. AIM 3 will test whether 5d-MI networks are engaged in motor plans based on cues that predict the future behavior of the visual stimulus, or about movement based upon internal representations. These experiments will test whether MI -5d networks use predictive visual cues to construct plans for upcoming movement and whether these networks compute motions for learned movement sequences. The proposed experiments will provide new information about corticocortical comp'Utations, including their serial and parallel processing and their coordinate frames. They will also help to advance multielectrode recording technologies and methods for statistical evaluation of single neuron and population coding. Finally, the results of these experiments will be useful in designing neural prosthetic devices to restore movement in paralyzed humans. 3ERFOFIMANCE SITE(S) (organization, city, state) Department of Neuroscience, Brown University, Providence, RI KEY PERSONNEL. See instructions on Page 11. Usecontinuationpages as neededto provide the required information in the format shown below. Name Organization Role on Project John Donoghue Lucien Bienenstock Michael Black Juliana Dushanova Department of Neuroscience,Brown University PrincipalInvestigator Department of Neuroscience, Brown University Co-Investigator Computer Science Dept., Brown University Co-Investigator Department of Neuroscience, Brown University Research Associate PHS391)(Rev. 5/95) Page 2 B B Number sages consecutively at the bottom throughout the application. Do not use suffixes such as 3a, 3b. CC PrindmPivestigator/Program Director (Last, first, middle): _jjPongghug. _JQ_hn Type the name of the principal investigator/program director at the top of each printed page and each continuation page. (For type specifications, see instructions on page6.) RESEARCH GRANT TABLE OF CONTENTS Page Numbers Face Page 1 Description,